A mixed methods exploration of the emergence of goal orientation in medical students' individualized learning plans.

PURPOSE: Goal orientation (GO) describes an individual's approach to different achievement situations. Understanding the motivations and approach to achieving goals of medical students is vital with the increasing emphasis on self-directed learning. The purpose of this study was to identify themes in self-improvement reflections that relate to each GO dimension (learning, performance-prove, and performance-avoid). METHODS: A sequential explanatory mixed methods design was used. GO data was used to categorize students into groups aligning with the GO dimension identified in the previous stage of quantitative analysis. Individualized learning plans (ILPs) for each GO dimension group were coded inductively to identify emergent themes related to goal setting and achievement. RESULTS: The learning GO group was the largest of the three GOs. Five themes were identified from inductive analysis: importance of practice, identifying elements that helped, identifying structural barriers, opportunities for improvement, and acknowledging experience. While these themes occur across GO, patterns exist within their ILPs based on GO. CONCLUSIONS: We identified common themes for motivations of medical students, and these motivations might differ depending on their GO. Further exploration into the themes over the course of their training will provide additional insights on what factors may be involved in student motivations towards learning and achievement. Educators can use this information to individualize feedback and students can better understand their motivations towards achieving goals.

What is the association between student well-being and high-stakes examination scores?

INTRODUCTION: As educators seek to improve medical student well-being, it is essential to understand the interplay between distress and important outcomes. Performance on Step 1 of the United States Medical Licensing Examination has played a significant role in selection for postgraduate residency positions in the United States and consequently has been a source of great stress for medical students. The purpose of this study was to examine whether student well-being correlates with performance on a high stakes licensing examination. METHODS: Between 2014 and 2016, three sequential cohorts of medical students at the University of Michigan Medical School completed the Medical Student Well-Being Index (MSWBI) at the end of their 2nd-year curriculum, shortly before taking Step 1. Associations between well-being and Step 1 scores were investigated while adjusting for MCAT scores and cumulative second-year course scores. RESULTS: In total, 354 students were included in the analysis (68.1% of potential responders). On bivariate analysis, poor student well-being (0 = low distress [high well-being], 7 = high distress [poor well-being]) was associated with lower Step 1 examination scores (slope = -2.10, P < .01), and well-being accounted for 5% of overall Step 1 score variability (R2  = .05). However, after adjustment for MCAT scores and cumulative GPA (full model R2  = .51), the relationship between well-being and Step 1 score was no longer significant (slope = -0.70, P-value = .06). CONCLUSIONS: When controlling for metrics of academic performance, student well-being prior to taking Step 1 was not associated with how well students performed on Step 1 for the study sample.

Beyond advising and mentoring: Competencies for coaching in medical education.

BACKGROUND: Coaching supports academic goals, professional development and wellbeing in medical education. Scant literature exists on training and assessing coaches and evaluating coaching programs. To begin filling this gap, we created a set of coach competencies for medical education using a modified Delphi approach. METHODS: An expert team assembled, comprised of seven experts in the field of coaching. A modified Delphi approach was utilized to develop competencies. RESULTS: Fifteen competencies in five domains resulted: coaching process and structure, relational skills, coaching skills, coaching theories and models, and coach development. CONCLUSION: These competencies delineate essential features of a coach in medical education. Next steps include creating faculty development and assessment tools for coaching.

Measuring Coaching in Undergraduate Medical Education: the Development and Psychometric Validation of New Instruments.

BACKGROUND: Coaching is emerging as a novel approach to guide medical students toward becoming competent, reflective physicians and master adaptive learners. However, no instruments currently exist to measure academic coaching at the undergraduate medical education level. OBJECTIVE: To describe the development and psychometric assessments of two instruments designed to assess academic coaching of medical students toward creating a robust measurement model of this educational paradigm. DESIGN: Observational psychometric. PARTICIPANTS: All medical students in the 2014 and 2015 cohorts and all their coaches were invited to complete the instruments being tested, which led to 662 medical student responses from 292 medical students and 468 coaching responses from 22 coaches being included in analyses. Medical student response rates were 75.7% from 2014 and 75.5% from 2015. Overall, the coaches' response rate was 71%. MAIN MEASURES: Two 31-item instruments were initially developed, one for medical students to assess their coach and one for faculty coaches to assess their students, both of which evaluated coaching based on definitions we formulated using existing literature. Each was administered to two cohorts of medical students and coaches in 2015 and 2016. An exploratory factor analysis using principal component analysis as the extraction method and Varimax with Kaiser normalization as the rotation method was conducted. KEY RESULTS: Eighteen items reflecting four domains were retained with eigenvalues higher than 1.0 for medical students' assessment of coaching, which measured promoting self-monitoring, relationship building, promoting reflective behavior, and establishing foundational ground rules. Sixteen items reflecting two domains were retained for the faculty assessment of coaching with eigenvalues higher than 1.0: the Practice of Coaching and Relationship Formation. CONCLUSIONS: We successfully developed and psychometrically validated surveys designed to measure key aspects of the coaching relationship, coaching processes, and reflective outcomes. The new validated instruments offer a robust measurement model for academic coaching.

Competencies for Those Who Coach Physicians: A Modified Delphi Study.

The rapidly evolving coaching profession has permeated the health care industry and is gaining ground as a viable solution for addressing physician burnout, turnover, and leadership crises that plague the industry. Although various coach credentialing bodies are established, the profession has no standardized competencies for physician coaching as a specialty practice area, creating a market of aspiring coaches with varying degrees of expertise. To address this gap, we employed a modified Delphi approach to arrive at expert consensus on competencies necessary for coaching physicians and physician leaders. Informed by the National Board of Medical Examiners' practice of rapid blueprinting, a group of 11 expert physician coaches generated an initial list of key thematic areas and specific competencies within them. The competency document was then distributed for agreement rating and comment to over 100 stakeholders involved in physician coaching. Our consensus threshold was defined at 70% agreement, and actual responses ranged from 80.5% to 95.6% agreement. Comments were discussed and addressed by 3 members of the original group, resulting in a final model of 129 specific competencies in the following areas: (1) physician-specific coaching, (2) understanding physician and health care context, culture, and career span, (3) coaching theory and science, (4) diversity, equity, inclusion, and other social dynamics, (5) well-being and burnout, and (6) physician leadership. This consensus on physician coaching competencies represents a critical step toward establishing standards that inform coach education, training, and certification programs, as well as guide the selection of coaches and evaluation of coaching in health care settings.

Competency-Based Medical Education in a Norm-Referenced World: A Root Cause Analysis of Challenges to the Competency-Based Paradigm in Medical School.

Competency-based medical education (CBME) requires a criterion-referenced approach to assessment. However, despite best efforts to advance CBME, there remains an implicit, and at times, explicit, demand for norm-referencing, particularly at the junction of undergraduate medical education (UME) and graduate medical education (GME). In this manuscript, the authors perform a root cause analysis to determine the underlying reasons for continued norm-referencing in the context of the movement toward CBME. The root cause analysis consisted of 2 processes: (1) identification of potential causes and effects organized into a fishbone diagram and (2) identification of the 5 whys. The fishbone diagram identified 2 primary drivers: the false notion that measures such as grades are truly objective and the importance of different incentives for different key constituents. From these drivers, the importance of norm-referencing for residency selection was identified as a critical component. Exploration of the 5 whys further detailed the reasons for continuation of norm-referenced grading to facilitate selection, including the need for efficient screening in residency selection, dependence upon rank-order lists, perception that there is a best outcome to the match, lack of trust between residency programs and medical schools, and inadequate resources to support progression of trainees. Based on these findings, the authors argue that the implied purpose of assessment in UME is primarily stratification for residency selection. Because stratification requires comparison, a norm-referenced approach is needed. To advance CBME, the authors recommend reconsideration of the approach to assessment in UME to maintain the purpose of selection while also advancing the purpose of rendering a competency decision. Changing the approach will require a collaboration between national organizations, accrediting bodies, GME programs, UME programs, students, and patients/societies. Details are provided regarding the specific approaches required of each key constituent group.

Coaching models, theories, and structures: An overview for teaching faculty in the emergency department and educators in the offices.

Coaching is rapidly advancing in medical education as a relational process of facilitating sustainable change and growth. Coaching can support learners in emergency medicine at any stage by improving self-reflection, motivation, psychological capital, and goal creation and attainment. Different from the traditional models of advising and mentoring, coaching may be a new model for many educators. An introduction to key coaching concepts and ways they may be implemented in emergency medicine is provided. Experienced coaches employ a diverse array of models and techniques that may be new to novice coaches. This article summarizes a variety of coaching models, theories, and content areas that can be adapted to a coachee's needs and the situational context-be it the fast-paced emergency department or the faculty member's office.

Leveling the field: Development of reliable scoring rubrics for quantitative and qualitative medical education research abstracts.

BACKGROUND: Research abstracts are submitted for presentation at scientific conferences; however, criteria for judging abstracts are variable. We sought to develop two rigorous abstract scoring rubrics for education research submissions reporting (1) quantitative data and (2) qualitative data and then to collect validity evidence to support score interpretation. METHODS: We used a modified Delphi method to achieve expert consensus for scoring rubric items to optimize content validity. Eight education research experts participated in two separate modified Delphi processes, one to generate quantitative research items and one for qualitative. Modifications were made between rounds based on item scores and expert feedback. Homogeneity of ratings in the Delphi process was calculated using Cronbach's alpha, with increasing homogeneity considered an indication of consensus. Rubrics were piloted by scoring abstracts from 22 quantitative publications from AEM Education and Training "Critical Appraisal of Emergency Medicine Education Research" (11 highlighted for excellent methodology and 11 that were not) and 10 qualitative publications (five highlighted for excellent methodology and five that were not). Intraclass correlation coefficient (ICC) estimates of reliability were calculated. RESULTS: Each rubric required three rounds of a modified Delphi process. The resulting quantitative rubric contained nine items: quality of objectives, appropriateness of methods, outcomes, data analysis, generalizability, importance to medical education, innovation, quality of writing, and strength of conclusions (Cronbach's α for the third round = 0.922, ICC for total scores during piloting = 0.893). The resulting qualitative rubric contained seven items: quality of study aims, general methods, data collection, sampling, data analysis, writing quality, and strength of conclusions (Cronbach's α for the third round = 0.913, ICC for the total scores during piloting = 0.788). CONCLUSION: We developed scoring rubrics to assess quality in quantitative and qualitative medical education research abstracts to aid in selection for presentation at scientific meetings. Our tools demonstrated high reliability.

Local media influence on opting out from an exception from informed consent trial.

STUDY OBJECTIVE: News media are used for community education and notification in exception from informed consent clinical trials, yet their effectiveness as an added safeguard in such research remains unknown. We assessed the number of callers requesting opt-out bracelets after each local media report and described the errors and content within each media report. METHODS: We undertook a descriptive analysis of local media trial coverage (newspaper, television, radio, and Web log) and opt-out requests during a 41-month period at a single site participating in an exception from informed consent out-of-hospital trial. Two nontrial investigators independently assessed 41 content-based media variables (including background, trial information, graphics, errors, publication information, and assessment) with a standardized, semiqualitative data collection tool. Major errors were considered serious misrepresentation of the trial purpose or protocol, whereas minor errors included misinformation unlikely to mislead the lay reader about the trial. We plotted the temporal relationship between opt-out bracelet requests and media reports. Descriptive information about the news sources and the trial coverage are presented. RESULTS: We collected 39 trial-related media reports (33 newspaper, 1 television, 1 radio, and 4 blogs). There were 13 errors in 9 (23%) publications, 7 of which were major and 6 minor. Of 384 requests for 710 bracelets, 310 requests (80%) occurred within 4 days after trial media coverage. Graphic timeline representation of the data suggested a close association between media reports about the trial and requests for opt-out bracelets. CONCLUSION: According to results from a single site, local media coverage for an exception from informed consent clinical trial had a substantial portion of errors and appeared closely associated with opt-out requests.

The AAMC Standardized Video Interview: Lessons Learned From the Residency Selection Process.

Calls to change the residency selection process have increased in recent years, with many focusing on the need for holistic review and alternatives to academic metrics. One aspect of applicant performance to consider in holistic review is proficiency in behavioral competencies. The Association of American Medical Colleges (AAMC) developed the AAMC Standardized Video Interview (SVI), an online, asynchronous video interview that assesses applicants' knowledge of professionalism and their interpersonal and communication skills. The AAMC worked with the emergency medicine community to pilot the SVI. Data from 4 years of research (Electronic Residency Application Service [ERAS] 2017-2020 cycles) show the SVI is a reliable, valid assessment of these behavioral competencies. It provides information not available in the ERAS application packet, and it does not disadvantage individuals or groups. Yet despite the SVI's psychometric properties, the AAMC elected not to renew or expand the pilot in residency selection.In this Invited Commentary, the authors share lessons learned from the AAMC SVI project about introducing a new tool for use in residency selection. They recommend that future projects endeavoring to find ways to support holistic review engage all stakeholders from the start; communicate the value of the new tool early and often; make direct comparisons with existing tools; give new tools time and space to succeed; strike a balance between early adopters and broad participation; help stakeholders understand the limitations of what a tool can do; and set clear expectations about both stakeholder input and pricing. They encourage the medical education community to learn from the SVI project and to consider future partnerships with the AAMC or other specialty organizations to develop new tools and approaches that prioritize the community's needs. Finding solutions to the challenges facing residency selection should be a priority for all stakeholders.

The Emergency Medicine Group Standardized Letter of Evaluation as a Workplace-based Assessment: The Validity Is in the Detail.

INTRODUCTION: Interest is growing in specialty-specific assessments of student candidates based on clinical clerkship performance to assist in the selection process for postgraduate training. The most established and extensively used is the emergency medicine (EM) Standardized Letter of Evaluation (SLOE), serving as a substitute for the letter of recommendation. Typically developed by a program's leadership, the group SLOE strives to provide a unified institutional perspective on performance. The group SLOE lacks guidelines to direct its development raising questions regarding the assessments, processes, and standardization programs employ. This study surveys EM programs to gather validity evidence regarding the inputs and processes involved in developing group SLOEs. METHODS: A structured telephone interview was administered to assess the input data and processes employed by United States EM programs when generating group SLOEs. RESULTS: With 156/178 (87.6%) of Accreditation Council of Graduate Medical Education-approved programs responding, 146 (93.6%) reported developing group SLOEs. Issues identified in development include the following: (1) 84.9% (124/146) of programs limit the consensus process by not employing rigorous methodology; (2) several stakeholder groups (nurses, patients) do not participate in candidate assessment placing final decisions at risk for construct under-representation; and (3) clinical shift assessments don't reflect the task-specific expertise of each stakeholder group nor has the validity of each been assessed. CONCLUSION: Success of the group SLOE in its role as a summative workplace-based assessment is dependent upon valid input data and appropriate processes. This study of current program practices provides specific recommendations that would strengthen the validity arguments for the group SLOE.

Emergency medical service providers' attitudes and experiences regarding enrolling patients in clinical research trials.

OBJECTIVE: The purpose of this study was to evaluate Emergency Medical Services (EMS) providers' attitudes and experiences about enrolling patients in clinical research trials utilizing the federal rules for exception from informed consent. We hypothesized that Emergency Medical Technicians (EMTs) would have varied attitudes about research using an exception from informed consent which could have an impact on the research. METHODS AND SETTING: Since January 2007, the EMS system has been participating in a randomized, multi-center interventional trial in which out-of-hospital providers enroll critically injured trauma patients using exception from informed consent.A voluntary, anonymous, written survey was administered to EMS providers during an in-service. The survey included demographics and Likert-type questions about their experiences with and attitudes towards research in general, and research using an exception from informed consent for an out-of-hospital clinical trial. RESULTS: The response rate was 79.3% (844/1067). Most respondents, 93.3%, agreed that "research in EMS care is important." However, 38.5% also agreed that individual EMTs/paramedics should maintain the personal right of refusal to enroll patients in EMS trials. Fifty-four percent of respondents agreed with the statement that "the right of research subjects to make their own choices is more important than the interests of the general community." In response to statements about the current study, 11.3% agreed that "the study is unethical because the patient cannot consent" and 69.2% responded that they would personally be willing to be enrolled in the study before they were able to give consent if they were seriously injured. Those who had not enrolled a patient into the study (681 respondents) were asked their reasons: 76.8% had not encountered an eligible patient or did not work for an agency that carried the fluid; 4.3% did not have time; 4.1% forgot and 1.1% stated that they were opposed to enrolling patients in studies without their consent. CONCLUSION: The majority of EMS personnel in one community support EMS research and this specific out-of-hospital clinical trial being conducted under an exception from informed consent. Potential barriers to enrollment were identified. Further study in other systems is warranted to better understand EMS provider perspectives about exception from informed consent research.

Differences in Coaching Needs Among Underrepresented Minority Medical Students.

BACKGROUND AND OBJECTIVES: Little is known about how the academic coaching needs of medical students differ between those who are racially, ethnically, and socially underrepresented minority (RES-URM) and those who represent the majority. This single-site exploratory study investigated student perceptions and coaching needs associated with a mandatory academic coaching program, and coaches' understanding of and preparedness to address these potentially differing needs. METHODS: Coaching needs of second- and third-year medical students were assessed using two initial focus groups and two validation focus groups, one consisting of RES-URM students and the other majority medical students. Coaches were assessed using a cross-sectional self-administered survey designed to determine their perceptions of differing coaching needs of students Results: Seven themes emerged from the student focus groups. Three of these reflected the coaching relationship, and four reflected the coaching process. RES-URM students expressed stress around sharing vulnerability that was not expressed among majority students. Sixty-eight percent of coaches expressed that RES-URM students would not have differing needs of their coaches. Coaches self-rated as being somewhat (45%), moderately (29%), or very (13%) skilled at coaching RES-URM students. CONCLUSIONS: RES-URM students cite different coaching needs than majority students that most coaches do not recognize. Faculty and program development regarding these unique needs is warranted.

Innovation in Residency Selection: The AAMC Standardized Video Interview.

PURPOSE: Innovative tools are needed to shift residency selection toward a more holistic process that balances academic achievement with other competencies important for success in residency. The authors evaluated the feasibility of the AAMC Standardized Video Interview (SVI) and evidence of the validity of SVI total scores. METHOD: The SVI, developed by the Association of American Medical Colleges, consists of six questions designed to assess applicants' interpersonal and communication skills and knowledge of professionalism. Study 1 was conducted in 2016 for research purposes. Study 2 was an operational pilot administration in 2017; SVI data were available for use in residency selection by emergency medicine programs for the 2018 application cycle. Descriptive statistics, correlations, and standardized mean differences were used to examine data. RESULTS: Study 1 included 855 applicants; Study 2 included 3,532 applicants. SVI total scores were relatively normally distributed. There were small correlations between SVI total scores and United States Medical Licensing Examination Step exam scores, Alpha Omega Alpha Honor Medical Society membership, and Gold Humanism Honor Society membership. There were no-to-small group differences in SVI total scores by gender and race/ethnicity, and small-to-medium differences by applicant type. CONCLUSIONS: Findings provide initial evidence of the validity of SVI total scores and suggest that these scores provide different information than academic metrics. Use of the SVI, as part of a holistic screening process, may help program directors widen the pool of applicants invited to in-person interviews and may signal that programs value interpersonal and communication skills and professionalism.

Applicant Reactions to the AAMC Standardized Video Interview During the 2018 Application Cycle.

PURPOSE: This study examined applicant reactions to the Association of American Medical Colleges Standardized Video Interview (SVI) during its first year of operational use in emergency medicine (EM) residency program selection to identify strategies to improve applicants' SVI experience and attitudes. METHOD: Individuals who self-classified as EM applicants applying in the Electronic Residency Application Service 2018 cycle and who completed the SVI in summer 2017 were invited to participate in 2 surveys. Survey 1, which focused on procedural issues, was administered immediately after SVI completion. Survey 2, which focused on applicants' SVI experience, was administered in fall 2017, after SVI scores were released. RESULTS: The response rates for surveys 1 and 2 were 82.3% (2,906/3,532) and 58.7% (2,074/3,532), respectively. Applicant reactions varied by aspect of the SVI studied and their SVI total scores. Most applicants were satisfied with most procedural aspects of the SVI, but most applicants were not satisfied with the SVI overall or with their total SVI scores. About 20% to 30% of applicants had neutral opinions about most aspects of the SVI. Negative reactions to the SVI were stronger for applicants who scored lower on the SVI. CONCLUSIONS: Applicants had generally negative reactions to the SVI. Most were skeptical of its ability to assess the target competencies and its potential to add value to the selection process. Applicant acceptance and appreciation of the SVI will be critical to the SVI's acceptance by the graduate medical education community.

The AAMC Standardized Video Interview: Reactions and Use by Residency Programs During the 2018 Application Cycle.

PURPOSE: To evaluate how emergency medicine residency programs perceived and used Association of American Medical Colleges (AAMC) Standardized Video Interview (SVI) total scores and videos during the Electronic Residency Application Service 2018 cycle. METHOD: Study 1 (November 2017) used a program director survey to evaluate user reactions to the SVI following the first year of operational use. Study 2 (January 2018) analyzed program usage of SVI video responses using data collected through the AAMC Program Director's Workstation. RESULTS: Results from the survey (125/175 programs; 71% response rate) and video usage analysis suggested programs viewed videos out of curiosity and to understand the range of SVI total scores. Programs were more likely to view videos for attendees of U.S. MD-granting medical schools and applicants with higher United States Medical Licensing Examination Step 1 scores, but there were no differences by gender or race/ethnicity. More than half of programs that did not use SVI total scores in their selection processes were unsure of how to incorporate them (36/58; 62%) and wanted additional research on utility (33/58; 57%). More than half of programs indicated being at least somewhat likely to use SVI total scores (55/97; 57%) and videos (52/99; 53%) in the future. CONCLUSIONS: Program reactions on the utility and ease of use of SVI total scores were mixed. Survey results indicate programs used the SVI cautiously in their selection processes, consistent with AAMC recommendations. Future user surveys will help the AAMC gauge improvements in user acceptance and familiarity with the SVI.

Community consultation methods in a study using exception to informed consent.

OBJECTIVE: The most effective means of community consultation is unknown. We evaluated differences in community opinion elicited by varying means of consultation. METHODS: We compared responses with a cross-sectional, standardized survey administered as part of the community consultation for the Resuscitation Outcomes Consortium (ROC) hypertonic saline trial. Surveys were obtained from four sources: two sets of random-digit dialing phone surveys, paper surveys from community meetings, and web-based surveys. RESULTS: Three hundred sixty-one usable surveys were obtained: 186 from phone survey 1; 86 from phone survey 2 (using slightly modified wording); 54 from community meetings (8 from open forums; 46 from existing meetings); and 35 from a web site. Demographics were similar between the sets except that the surveys obtained from community meetings had the highest minority representation (63.3% nonwhite). Community meeting respondents were more willing than phone or web respondents to receive experimental treatment for themselves (93.6% vs. 77.5% overall) and for a family member (95.2% vs. 74.9% overall). The web-based survey generated the least feedback and had the most higher-income responders. CONCLUSIONS: Responses varied by method of consultation. The open forums were very poorly attended, despite heavy advertising by investigators. Furthermore, attendees at those meetings provided the least objection to proposed research without informed consent. Phone surveys elicited the most objections. We suggest that an efficient method of community consultation is random-digit dialing supplemented with discussion at already scheduled events to target special populations.

Specialty Choice Stability: Are There Implications for Early Entry Into Residency?

INTRODUCTION: The medical education community is piloting early entry to residency as a potential outcome to competency-based education and as a way to more quickly train future physicians in specialties of need. However, there is limited knowledge of which specialties may be best suited to this streamlined training. Student career desires may change over the course of their undergraduate training. We aimed to understand which specialties have stable student interest from matriculation to residency match in an effort to highlight which fields may be reasonable to consider for such accelerated programs. METHODS: Medical students at one school of medicine who matriculated in the years 2009-2013 were surveyed upon entry regarding the medical specialty they were most interested in pursuing. Six hundred fifty-four students were eligible for inclusion and 535 of the records met all requirements. On completion of medical school, final specialty choice for students obtaining a residency position was recorded. The data were analyzed to assess specialties with the highest versus the lowest rates of retention. RESULTS: Of 535 included students, the top specialties with retention of interest (no change in specialty choice for that student) from matriculation to match were physical medicine and rehabilitation, (100%, n=3 retention/3 initial), psychiatry (57.1%, 4/7), internal medicine (48.5%, 47/97), and family medicine (47.7%, 41/86). The specialties with the lowest retention were pathology (0%, 0/2), preventive medicine (0%, 0/4), dermatology (12%, 1/8), neurology (16.7%, 3/18) and radiation oncology (16.7%, 1/6). DISCUSSION: Some specialties that attract student interest before matriculation may be more likely to maintain interest compared to others. This suggests a need for further research to determine if residency entry can begin earlier than traditionally thought, with certain fields better suited for accelerated training.